Welcome to our third online lecture in the Fall 2015 Introduction to Sociology course. This week’s lecture takes on two tasks. On the one hand, we’ll consolidate and review materials in the course so far, particularly the course syllabus, our academic integrity policy, and questions from last lecture that challenged our common sense. On the other hand, we’ll look forward. If the subject of sociology is the study of structures built on interaction of individuals, groups and institutions, and if the stories we tell about that interaction are divided into four rough camps called paradigms, then the point of all this paradigm-building and studying is to better understand the real world (where “realness” has to do with consistent, observable consequences). Sociologists build their understanding through research, and it is the basic nature of sociological research that we contemplate this week.
Thinking Like a Researcher: Questions, Theories, Hypotheses, and Variables
Not all Questions are Sociological Research Questions
“Question Reality,” reads the famous t-shirt slogan. But there are many ways to question reality, and not all questions are good sociological research questions. “Why does Frederica hate tofu?” isn’t a very good research question, because it’s narrow in scope, limited to Frederica alone. Good research questions are broad enough to cover a wide variety of people.
“Why is all American cash printed green?” is a broad question, but it isn’t a very good sociological research question because it doesn’t describe a variable. A variable is a characteristic that can be measured and that has different observed values. Because American cash only comes printed green, its color is not a variable. Good research questions describe variables, not constants.
“Where are the invisible signs that point toward the undisclosed location of the UFO base?” indicates a variable (sign location), but it isn’t a very good research question because its variable isn’t observable. If signs can’t be seen, and the location of the supposed UFO base has not been disclosed, the research question simply can’t be answered through observation. Good research questions describe variables that can be observed.
“Is a higher level of educational attainment associated with higher income?” is a good research question because it is broad enough to cover a wide range of people and because it contains variables (educational attainment and income) that can be observed. The variable educational attainment can be observed in a number of ways: highest grade passed in school, perhaps, or highest degree obtained. The variable income is usually measured in terms of currency obtained over a period of time.
Hypotheses Connect Theories to Variables
The answer to a research question is a hypothesis, a prediction describing the direction of a relationship between two variables. The relationship between two variables can take on three possible directions:
A positive relationship between variables indicates that as the value of one variable goes up, the value of the other variable goes up as well. The hypothesis “income rises as educational attainment rises” predicts a positive relationship.
A negative relationship occurs when a rise in one variable is associated with a fall in the other variable. The hypothesis “income falls as educational attainment rises” predicts a negative relationship.
No relationship occurs when a rise in the value of one variable leads to no change in the other variable. The hypothesis “income does not change when educational attainment rises” predicts no relationship.
Finally, hypotheses commonly specify an independent variable and a dependent variable. Dependent variables describe an outcome that depends on the value of the independent variable. The value of an independent variable, on the other other hand, typically does not rely on the value of the dependent variable. To provide an example, we would say that when describing the relationship between the variables “presence of daylight” and “air temperature,” air temperature is the dependent variable while presence of daylight is an independent variable. Why? Because the value of air temperature depends on whether daylight is present. On the other hand, the presence of daylight is not influenced at all by air temperature, so it is not a dependent variable in the relationship between the two variables — it is an independent variable.
The example of daylight and air temperature is an easy one, because much observation has established the effect of one variable upon the other. Those engaged in sociological research, however, are often examining sociological phenomena for the first time. What tells us which variable is the independent variable and which is the dependent variable? Theory, the story we tell about how the social world works.
When it comes to educational attainment and income, what’s your theory? Do you believe that earning a higher education leads to higher income because advanced degrees qualify people for better-paying jobs? If so, then you would propose a hypothesis in which educational attainment is an independent variable and income is the dependent variable. Do you believe, on the other hand, that earning a higher income leads to better educational attainment because it takes money to pay for a college degree? If so, then you would propose a hypothesis in which educational attainment is the dependent variable and income is the independent variable.
Data Connect Hypotheses to the Observable World
Prediction takes us only so far. At some point, we need to compare our predictions against observations. This is when we begin collecting data. Every target for observation is called a case, and a dataset is built up by gathering together measurement of the same variables across many cases.
The example below shows a typical organization for a dataset. Every case is described by first name, and a unique case id number is also assigned, just in case two people in the dataset have the same name. Every case has its own row, and every variable has its own column.
Once we get down to the practical level of collecting data, every variable must be operationalized. That is, the researcher must describe exactly how a variable will be measured. Income is a vague term when you think about it. It could refer to an hourly wage, or the size of a biweekly check. In the table below, we’ll operationalize income as dollars earned per year. And what about educational attainment? A typical operationalization is to ask a person to name the highest-rank educational degree they have earned.
|Case ID||Name||Variable: Annual Income ($ per year)||Variable: Highest Degree Earned|
|2||Ernie||$40,000||High School Diploma|
|4||Frank||$14,400||High School Diploma|
|6||Nancy||$30,000||High School Diploma|
The way income has been operationalized is fundamentally different from the way educational attainment has been operationalized. Income is measured as a numerical variable — in other words, as a number. Educational attainment is not measured numerically, but rather as a categorical variable — a series of categories. We could have operationalized income categorically, into “high income,” “middle income,” “low income” and “no income,” for instance. We also could have operationalized educational attainment numerically, according to grade or grade-equivalent (a B.A. would be 16th grade, a Master’s would be 18th grade, and so on).
Once we’ve gathered our short and simple dataset, we can answer our research question, “Is a higher level of educational attainment associated with higher income?” To obtain our answer, let’s calculate the average income for different levels of educational attainment.
- Average Income, High School Diploma: $28,133
- Average Income, Bachelor’s Degree: $52,500
- Average Income, Master’s Degree: $55,000
- Average Income, PhD: $80,000
In this example, a positive relationship between educational attainment and income is apparent.
The most common method for creating a table like the one you see above is to use a spreadsheet program. The spreadsheet program Google Sheets is associated with your university account and you can use it for free. In this week’s in-class activity, Parking Lot Sociology, I’ll be asking you to collect data, enter it into a Google Sheets spreadsheet, describe the relationship between two variables, then share the results with me. Watch the two videos below; the second video shows you exactly how to use Google Sheets to get your work done.
An Example in Measuring Variables: By the Book
The first, short video segment below is based on the premise that you can find variables just about anywhere — even in a bookshelf. Looking at a bookshelf, it’s possible to identify three kinds of cases you might observe: the book, the shelf, and book pairs. How are books like people?
This second video is an important walkthrough exercise, demonstrating the skills you’ll need to complete this week’s in-course activity, Parking Lot Sociology:
The video above provides a shortened example of the kind of work you’ll be carrying out in your Parking Lot Sociology activity, but it should be enough to get you ready for this week’s activity. We’ll be carrying out Parking Lot Sociology during class time, so if you have questions after these videos, bring them and be ready to ask them!
Hirschi, Travis, and Michael Gottfredson. 1985. “Age and crime, logic and scholarship: Comment on Greenberg.” American of Journal Sociology 91(1): 22-27.
Schulz, Wolfram, John Ainley, Julian Fraillon, David Kerr, and Bruno Losito. 2010. ICCS 2009 International Report: Civic Knowledge, Attitudes, and Engagement among Lower-Secondary School Students in 38 Countries. International Association for the Evaluation of Educational Achievement. Amsterdam, the Netherlands: ICCS.
University of Maine at Augusta. 2015. Student Academic Integrity Code. Accessed January 20, 2015 at http://www.uma.edu/studenthandbookpol-s.html#saic.
Vandehey, Michael, George Diekhoff and Emily LaBeff. 2007. “College Cheating: A Twenty-Year Follow-Up and the Addition of an Honor Code.” Journal of College Student Development 48: 468-480.